Protection within SONET Networks
Synchronous Optical NETwork (SONET) and Synchronous Digital Hierarchy (SDH) based networks typically emphasize redundancy. That is for example, should a particular network line that couples a pair of networking systems (which may also be referred to as “nodes”) within the network fail (or degrade), the network is designed to “switch over” to another network line so that traffic flow is not substantially interrupted. Various types of redundancy may be designed into a SONET network. Some examples are illustrated in the discussion that follows.
FIG. 1 shows a point-to-point perspective. Point to point redundancy focuses on the behavior of a pair of nodes 131, 132 that are coupled together by a plurality of SONET lines 1041, 1042, . . . 104x-1, 104x. Although other point-to-point schemes may be possible, common point-to-point schemes typically include 1+1 and 1:N. Both schemes classify a network line as either a working line or a protection line. A working line is deemed as the “active” line that carries the information transported by the network. A protection line serves as a “back-up” for a working line. That is, if a working line fails (or degrades), the protection line is used to carry the working line's traffic.
In a 1+1 scheme, both the working and protection lines simultaneously carry the same traffic. For example, referring to FIG. 1, if line 1041 is the working line and line 1042 is the protection line; the transmitting node 131 simultaneously transmits the same information on both the working line 1041 and the protection line 1042. The receiving node 132, during normal operation, “looks to” the working line 1041 for incoming traffic and ignores the incoming traffic on the protection line 1042. If a failure or degradation of the working line 1041 is detected, the receiving node 132 simply “looks to” the protection line 1042 for the incoming traffic (rather than the working line 1041).
In a 1:N scheme one protection line backs up N working lines (where N is an integer greater than or equal to 1). For example, referring to FIG. 1, lines 1041 through 104x-1 may be established as the working lines while line 104x may be established as the protection line. If any of the working lines 1041 through 104x-1 fail or degrade, the transmitting node 131 sends the traffic of the failed/degraded working line over the protection line 104x. The receiving node 132 also “looks to” the protection line 104x for the traffic that would have been sent over the failed/degraded working line prior to its failure/degradation.
FIG. 2 shows a ring perspective. Ring redundancy schemes focus on the behavior of a plurality of nodes 231 through 234 coupled together in a ring. Redundancy is commonly handled by sending identical streams of traffic in opposite directions. A first direction may be referred to as the working direction while a second direction may be referred to as the protection direction. The most commonly used types of ring protection are Unidirectional Path Switched Ring (UPSR) and Bi-directional Line Switched Ring (BLSR). In a two-fiber UPSR approach, working traffic is sent in a first direction around the ring (e.g., clockwise) and protection traffic is sent in a second direction around the ring (e.g., counter-clockwise).
In a Bi-directional Line Switched Ring (BLSR) approach, each pair of rings are viewed as comprising an “inner” ring and “outer” ring (although note that the rings' actual geographic coverage does not necessarily have to correspond to the inner ring always being within the outer ring). Typically, for each ring, half of the capacity is allocated for working traffic and the other half of the capacity is allocated for protection traffic. As such, both working traffic and protection traffic flow bi-directionally. In either the UPSR or BLSR approaches, if failure or degradation occurs in the working direction, active traffic is looked for in the protection direction.
More sophisticated SONET networks may also be designed that provide protection at higher degrees of resolution. That is, each SONET line (such as line 1041 of FIG. 1 or line 204 of FIG. 2) may be viewed as transporting a number of STS-1 signals. For example, if lines 1041 and 204 each correspond to an STS-n line, each of these lines may be viewed as carrying n STS-1 signals (e.g., an STS-48 line may be viewed as carrying 48 STS-1 signals). Furthermore, in other environments, each STS-1 signal is used as a resource for carrying a plurality of lower speed signals.
Protection may be provided for STS-1 signals individually or for their constituent lower speed signals individually. Either of these forms of protection are commonly referred to as “path protection”. For example, in one type of 1+1 path protection scheme, an individual “working” STS-1 signal within an STS-n line (rather than all the STS-1 signals on the STS-n line) is backed up by a “protection” STS-1 signal transported on another STS-n line.
Automatic Protection Switching (APS) is a protocol that may be executed by the nodes within a SONET network. APS allows SONET nodes to communicate and organize the switching over from their working configuration to a protection configuration in light of a failure or degradation event (and then back again after the failure/degradation is corrected). For example, in a typical approach, K1 and K2 bytes are embedded within the SONET frame that is communicated between a pair of nodes in order to communicate failure/degradation events, requests for a switch over, correction thereafter, etc.
Distributed Switch Architecture
FIG. 3 shows an embodiment of a distributed “full mesh” node (or system) architecture 331. The architecture 331 of FIG. 3 may be utilized to implement a SONET node such as nodes 131, 132 of FIG. 1 or nodes 231 through 234 of FIG. 2. An ingress channel receives incoming data from a networking line. FIG. 3 shows ingress channels 3011 through 301x that each receive incoming data on a respective network line 3031 through 303x.
An egress channel transmits outgoing data onto a networking line. FIG. 3 shows egress channels 3121 through 312x that each transmit outgoing data on a respective network line 3041 through 304x. In a full mesh architecture embodiment, each ingress channel 3011 through 301x transmits all of its ingress traffic to each egress channel 3121 through 312x. For example, referring to FIG. 3, ingress channel 3011 receives n STS-1 signals from its corresponding network line 3031 (e.g., if network line 3031 is an OC-48 line; n=48 and the ingress line channel receives 48 STS-1 signals).
All n of the STS-1 signals received by the ingress channel 3011 are transmitted across the node's backplane 305 over each of its output lines 306, 310, 311, 312. A backplane is a board (e.g., a PC or “planar” board) having signal lines that electrically couple various line cards together. Typically, individual cards “plug into” a backplane 305 (e.g., via a card connector) and; in so doing, become communicatively coupled with one another. As such, plugging a sufficient type and quantity of cards into a backplane results in the formation of a networking system. Frequently (although not a strict requirement), backplanes have little or no sophisticated circuitry (e.g., processors, Application Specific Integrated Circuits (ASICs), etc.) and are substantially a collection of “short circuits” from card to card (although passive devices (e.g., capacitors, resistors, etc.), line drivers and other signal enhancing chips or devices are often found on a typical backplane).
Continuing with the discussion of the full mesh embodiment of FIG. 3, note that each egress channel 3121 through 312x receives all n STS-1 signals received by ingress channel 3031. In one approach, each STS-1 signal is provided its own signal line to each egress channel. As a result, each output 306, 310, 311, 312 corresponds to a n-wide bus. As each ingress channel is similarly designed, each egress channel 3121 through 312x receives all the incoming traffic received by the node. For example, in the particular example of FIG. 3, there are x ingress channels 3011 through 301x that each receive n STS-1 signal. As such, each egress channel 3121 through 312x receives xn STS-1 signals (which correspond to the total amount of traffic received by the node 331).
For example, note that egress channel 3121 receives inputs 3061 through 306x where each of these inputs correspond to the n STS-1 signals received by their corresponding ingress channel (i.e., input 306 for ingress channel 3011, input 307 for ingress channel 3012, input 308 for ingress channel 3013, . . . and input 309 for ingress channel 301x). In order to implement the switching fabric of the node, each egress channel 3121 through 312x is configured to select n of the xn STS-1 signals and transmit the n STS-1 signals over its corresponding outgoing networking line 3041 through 304x.
Note that there is a distinction between a line card and a channel. A line card is a card that can be coupled to one or more network lines. A channel is a data path within a line card that handles traffic flow in a particular direction (e.g., ingress or egress). As such, a line card having both ingress and egress connectivity will have both an ingress channel and an egress channel. Thus, for example, the ingress channel 3011 and the egress channel 3121 may coexist upon the same line card.
In other distributed switch architecture embodiments, each ingress channel 3011 through 301x transmits less than all of its ingress traffic to each egress channel 3121 through 312x. For example, referring again to FIG. 3, if ingress channel 3011 receives n STS-1 signals from its corresponding network line 3031—less than all n of the STS-1 signals received by the ingress channel 3011 are transmitted across the node's backplane 305 over each of its output lines 306, 310, 311, . . . 312.
This reduced amount of backplane 305 traffic (as compared to the “full mesh” approach described above) may be achieved by designing some degree of switching intelligence into the ingress channels 3011 through 301x themselves. As a result, each ingress channel 3011 through 301x transmits a subset of all n of the STS-1 signals it receives. Note that the actual electrical signal lines used to transport STS-1 signals across the backplane 305 may also vary from embodiment (regardless of the backplane is a full mesh backplane or is less than a full mesh backplane). For example, in one approach STS-1 signals from the same frame may be transported in parallel (e.g., as suggested by FIG. 3) or in series (e.g., STS-1 signals from the same frame are transported in a time division multiplexed fashion).